Smoothed Bootstrap and Statistical Data Cloning for Classifier Evaluation

This work is concerned with the estimation of a classifier's accuracy. We first review some existing methods for error estimation, focusing on cross-validation and bootstrap, and motivate the use of kernel-based smoothing for small sample size. We use the term data cloning to refer to the process of (re)sampling the data via kernel-based smoothed bootstrap. A number of novel estimators based on cloning is presented. Finally, we extend our estimators to to allow cloning of complex real-life data sets, in which a data point may include continuous, bounded, integer and nominal attributes. This allows for better 1 

• Type: Conference
• Year: 2001
• Where: ICML
• Authors: Gregory Shakhnarovich, Ran El-Yaniv, Yoram Baram
• Keywords: Error Estimation, ICML 2001, Kernel-based Smoothed Bootstrap, Machine Learning, Real-life Data Sets